search

haskell / cabal-website

The http://www.haskell.org/cabal/ website
4 stars 15 forks source link

How does deployment work #39

Open BinderDavid opened 3 months ago

BinderDavid commented 3 months ago

The README.md says that the site is deployed from the gh-pages branch, but that does not seem to be correct anymore. The last time that branch was updated from a travis CI action was 4 years ago, but the published website contains changes that are newer than that. The site does not seem to be deployed automatically, because https://www.haskell.org/cabal/ does not contain the changes from https://github.com/haskell/cabal-website/pull/38

BinderDavid commented 3 months ago

Ah, I could already answer this myself: https://github.com/haskell/cabal-website/issues/5#issuecomment-1873766247 It is deployed manually. The issue came up here: https://discourse.haskell.org/t/ghc-and-cabal-the-big-picture/9968/20?u=davidb

ulysses4ever commented 3 months ago

Indeed, thanks for the ping!

@Mikolaj it'd be great to deploy the current version that fixed horribly outdated instructions (Haskell Platform etc.)

Mikolaj commented 3 months ago

Indeed, let's deploy. I remember that was a part of the release checklist. Did it vanish in the last release checklist purge? Or is the deployment process described elsewhere?

BinderDavid commented 3 months ago

I also discovered that the gh-pages branch is currently used to host the haddocks of Cabal-3.3, which is still reachable under https://haskell.github.io/cabal-website/doc/html/Cabal/ That doesn't seem to be very useful, since Cabal-3.3 is only a development version that was never released. Maybe we should just disable the github pages feature for this repo and delete the gh-pages branch as well?

ulysses4ever commented 3 months ago

Good catch. Should probably be disabled. @Mikolaj, @Kleidukos?

Mikolaj commented 3 months ago

Yes, let's gut it!

Mikolaj commented 3 months ago

Re deployment, it's described in https://github.com/haskell/cabal/wiki/Making-a-release#c6-publishing-everything-else, and I just did it and it works. Let me know if this needs an improvement or if we can/should limit the bus factor --- I have no idea who beside me has the right --- all haskell org members maybe (I don't think I'm a member so I must have gotten it from Gershom or somebody during my first releases/deployments)?